Acoustic Features Selection of Speaker Verification Based on Average KL Distance

Abstract:

This paper proposes a new Average Kullback-Leibler distance to make an optimal feature selection algorithm for the matching score fusion of speaker verification. The advantage of this novel distance is to overcome the shortcoming of the asymmetry of conventional Kullback-Leibler distance, which can ensure the accuracy and robustness of the computation of the information content between matching scores of two acoustic features. From the experimental results by a variety of fusion schemes, it is found that the matching score fusion between MFCC and residual phase gains most information content. It indicates this scheme can yield an excellent performance.

• Periodical: Applied Mechanics and Materials (Volumes 373-375)
• Pages: 629-633
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.373-375.629
• Online since: August 2013
• Authors: Yu Luan, Hong Zuo Li, Ya Fei Wang
• Keywords: Average Kullback-Leibler Distance, Feature Selection, Matching Score Fusion, Network Security, Speaker Verification

References

[1] NAKAGAWA S, Wang L, Ohtsuka S. Speaker Identification and Verification by Combining MFCC and Phase Information[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2012, 20(4): 1085-1095.

DOI: 10.1109/tasl.2011.2172422

[2] Martinez J, Perez H, Escamilla E, Suzuki M. Speaker Recognition Using Mel Frequency Cepstral Coefficients (MFCC) and Vector Quantization (VQ) Techniques[C]/ Proceedings of the 22nd International Conference on Electrical Communications and Computers (CONIELECOMP): 2012, 248-251.

DOI: 10.1109/conielecomp.2012.6189918

[3] Patil H A, Parhi K K. Development of TEO Phase for Speaker Recognition[C]/ Proceedings of the 2010 International Conference on Signal Processing and Communications (SPCOM): 2012, 1-5.

DOI: 10.1109/spcom.2010.5560486

[4] Djamel A, Boudraa B. A Robust Distributed Speech Recognition in Mobile Communications[C]/ Proceedings of the Developments in E-systems Engineering (DeSE), 2011, 479-483.

DOI: 10.1109/dese.2011.94

[5] Sant'Ana R, Coelho R, Alcaim A. Text-independent Speaker Reocognition based on the Hurst Parameter and the Multidimensional Fractional Brownian Motion Model[J]. IEEE Transaction on Audio, Speech and Language processing, 2006, 14(3): 931-940.

DOI: 10.1109/tsa.2005.858054

[6] Adachi Y, Kawamoto S, Morishima S, Nakamura S. Perceptual Similarity Measurement of Speech by Combination of Acoustic Features[C]/ Proceedings of the IEEE International Conference Acoustics, Speech, and Signal Processing: 2008, 4861-4864.

DOI: 10.1109/icassp.2008.4518746

[7] Liu D, Sun D, Qiu Z. Feature Selection for Fusion of Speaker Verification via Maximum Kullback-Leibler Distance[C]/ Proceedings of the 2010 IEEE 10th International Conference on Signal Processing (ICSP): 2010 , 565-568.

DOI: 10.1109/icosp.2010.5655871

[8] Campbell J P, Shen W, Campbell W M, Reva S, Jean-Franqois B, Driss M, Forensic Speaker Recognition[J]. IEEE Signal Processing Magazine, 2009, 26(2): 95 -103.

DOI: 10.1109/msp.2008.931100

[9] Wang L, Ohtsuka S, Nakagawa S. High Improvement of Speaker Identification and Verification by Combing MFCC and Phase Information[C]/ Proceedings of the IEEE International Conference Acoustics, Speech, and Signal Processing: 2009, 4529-4532.

DOI: 10.1109/icassp.2009.4960637

[10] The NIST 2001 Speaker ID Evaluation Protocol, [Online], Available: http: /www. nist. gov/ speech/tests/spk/2001/index. htm.

[11] Murty K, Yegnanarayana B. Combining Evidence from Residual Phase and MFCC Features for Speaker Recognition[J]. IEEE Signal Processing Letters, 2006, 13(1): 52-55.

DOI: 10.1109/lsp.2005.860538

[12] Guerin B, El Fakhri G. Realistic PET Monte Carlo Simulation With Pixelated Block Detectors, Light Sharing, Random Coincidences and Dead-Time Modeling[J]. IEEE Transactions on Nuclear Science, 2008, 55(3): 942- 952.

DOI: 10.1109/tns.2008.924064